Los Angeles heat waves haven't gotten hotter at their peak. Their nights have.

What counts as a heat wave here

There is no single temperature that means "heat wave" in Los Angeles: 80°F is a hot day at LAX and an ordinary one in Burbank. So the extreme is defined relative to each place's own climate.

That makes it 80°F at LAX and 95°F in Burbank. Nothing else is tuned — no humidity, no minimum night temperature, no adjustment for the season. What happens at night falls out of the data, not the definition. A day counts only when the hourly record covers it; a missing day breaks a run (how).

That identifies roughly 3.6 heat waves a summer at a typical station, covering about 16 days. (The threshold is the warmest 10% of May–October, so a station sees only about 18 threshold-crossing days a summer to begin with; requiring 3 in a row makes qualifying runs rarer still.) Stricter definitions change how many events qualify — 1.5 a summer at the 95th percentile, 0.5 a summer at the 98th — but not the central result: daytime peaks have changed little, while heat-wave nights are several degrees warmer (the table).

Every heat wave at six long-record LA-area stations

One line per heat wave, from its coolest night up to its hottest afternoon. Short bars mark the averages of both ends for 1951–1980 and the last 30 complete summers. Stations where the nights moved most relative to the afternoons come first; hollow dots are waves in a summer that is not yet complete.

Then vs now: three temperatures from every heat wave

For each of the 6 stations with a record in both windows: the average hottest afternoon of a heat wave, the average coolest heat-wave night (the lowest overnight temperature from the second night on — did the outdoor air ever substantially cool overnight?), and the night after the run broke. 1951–1980 against the last 30 complete summers.

95% intervals bootstrap entire summers, so several heat waves in one year aren't treated as independent observations; the pooled numbers resample stations as well (how).

Heat waves themselves aren't obviously longer or more frequent under this definition: 4.4 → 4.3 days per event (95%: −0.6 to 0.3) and 3.78 → 3.65 per summer (−2.25 to 1.40). The story is not that there are more of them; it is that when one happens, the night behaves differently. Year-by-year counts are in the explorer.

Overnight relief: hours under 70°F on a heat-wave night

A daily low of 69°F and one of 60°F both count as "under 70," but they are very different nights. The hourly readings can tell them apart: of the 14 hours between 6 pm and 8 am, how many were under 70°F? That number is closely tied to whether windows can be opened, whether a building sheds the day's heat, whether people sleep.

Station by station:

Share of 6 pm–8 am readings under 70°F, scaled to 14 hours so hourly and 3-hourly years compare; nights from the second night of the wave on.

The pattern isn't universal: March Air Reserve Base is the exception

March Air Reserve Base: the coolest heat-wave night moved +1.1° (−0.7° to +2.9°), and its ordinary summer nights moved +0.9°. At March, inland and away from the marine layer, heat-wave nights have changed by only about a degree — and so have ordinary summer nights. That is a real difference between places, not noise in the headline: this is a regional pattern with heterogeneous stations, not every thermometer proving one headline.

13 more stations have heat-wave records but too few mid-century summers to compare then with now. Their waves, same chart: Point Mugu, 59 complete summers.